Comentários (4)

1 LeRoiSoleil comentou às Link permanente

I read Topology Builder agents articles 1, 2 & 3. How should I interpret the Topology Agents Groups Status when I see (on Aug 26)
Integration Finished Aug 22, 4:41 PM
Dependency Finished Aug 23, 7:29 PM
Cleanup Finished Aug 23, 8:25PM
Background "No status" Aug 22, 3:51

Is the current topology ready for show time? Wait until Background finishes? Should I run Background manually?

2 LeRoiSoleil comentou às Link permanente

When I submit discovery of (a few) newly provisioned computers, will the topology builder agents complete more quickly than if the discovery comprised a large scope? The goal of a small scope discovery is to collect and display current information quickly. My question is the latency due to the topology builder agents.

3 jenlee comentou às Link permanente

Hi Ravi - I suspect that the Background group is still running. There are 6 agents in that group;

1. DatabaseServerCleanupAgent
2. HostDependencyAgent
3. MssObjLinkCleanupAgent
4. RuntimeCleanupAgent
5. SplitTableCleanupAgent
6. RegistrationInfoAgent
Look in dist/log/TADDM.log on the primary storage server and find the last occurrence of the word 'agent' and see if it is a start message (CTJOT0402I) for one of the above. If there is a start and no stop (CTJOT0403I) then the group is still running. If you see one of these agents running for a long time(more then a couple of hours), you may want to open a PMR for it.

4 jenlee comentou às Link permanente

Hi Ravi - in regards to your second comment, some agents will be quicker and some will not, depends on the scope of the agent. For example, the cleanup agents work on entire tables and use row counts to limit their work and are not scope specific. But many dependency agents are scoped by last modified time. When the agent starts, it checks the last time it ran(see table tb_agent_last_run) and only looks at CIs modified since that time. Not all dependency agents use this method, and as I get into the dependency group I will try to note when this is the case. So, short answer, yes, some agents will complete quicker on a small discovery then a large discovery. Keep in mind that the topology builder agents run on intervals, so generally items are up to date if you are discovering at regular intervals. Unless dependencies are changing a lot, or dependencies are very critical, there is probably not much changing from run to run that make the agents worth waiting for. I think it depends on your use case though.

Incluir um Comentário Incluir um Comentário